Silencing Threonine Deaminase and in Nicotiana attenuata ... › content › plantcell › 18 › 11...

Silencing Threonine Deaminase and JAR4 in Nicotianaattenuata Impairs Jasmonic Acid–Isoleucine–MediatedDefenses against Manduca sexta W

Jin-Ho Kang, Lei Wang, Ashok Giri, and Ian T. Baldwin1

Department of Molecular Ecology, Max-Planck-Institute of Chemical Ecology, D-07745 Jena, Germany

Threonine deaminase (TD) catalyzes the conversion of Thr to a-keto butyrate in Ile biosynthesis; however, its dramatic

upregulation in leaves after herbivore attack suggests a role in defense. In Nicotiana attenuata, strongly silenced TD

transgenic plants were stunted, whereas mildly silenced TD transgenic plants had normal growth but were highly

susceptible to Manduca sexta attack. The herbivore susceptibility was associated with the reduced levels of jasmonic acid–

isoleucine (JA-Ile), trypsin proteinase inhibitors, and nicotine. Adding [13C4]Thr to wounds treated with oral secretions

revealed that TD supplies Ile for JA-Ile synthesis. Applying Ile or JA-Ile to the wounds of TD-silenced plants restored

herbivore resistance. Silencing JASMONATE-RESISTANT4 (JAR4), the N. attenuata homolog of the JA-Ile–conjugating

enzyme JAR1, by virus-induced gene silencing confirmed that JA-Ile plays important roles in activating plant defenses. TD

may also function in the insect gut as an antinutritive defense protein, decreasing the availability of Thr, because continuous

supplementation of TD-silenced plants with large amounts (2 mmol) of Thr, but not Ile, increased M. sexta growth. However,

the fact that the herbivore resistance of both TD- and JAR-silenced plants was completely restored by signal quantities (0.6

mmol) of JA-Ile treatment suggests that TD’s defensive role can be attributed more to signaling than to antinutritive defense.

INTRODUCTION

Threonine deaminase (TD) catalyzes the formation of a-keto

butyrate (a-KB) from Thr, the first step in the biosynthesis of Ile.

Regulation of TD activity by Ile was the first recognized instance of

allosteric feedback regulation by the end product of a biosyn-

thetic pathway (Umbarger, 1956). The function of TD for Ile bio-

synthesis was demonstrated by analyzing the Ile auxotrophic

mutant in Nicotiana plumbaginifolia, which has no detectable TD

activity (Sidorov et al., 1981). When this mutant was transformed

with the Saccharomyces cerevisiae ILV gene that encodes TD, the

transformed lines could be grown on medium without Ile (Colau

et al., 1987). These results demonstrate that TD regulates Ile

production and is indispensable for plant growth. However, TD’s

unusual expression pattern in solanaceous plants suggests that

TD plays additional roles in development and herbivore defense.

For more than a decade, TD has been recognized as a reliable

marker for wounding and jasmonic acid (JA) elicitation in potato

(Solanum tuberosum) and tomato (Solanum lycopersicum)

(Hildmann et al., 1992; Samach et al., 1995; Dammann et al.,

1997). Wound-induced TD expression is mediated by abscisic

acid and JA signaling in tomato plants (Hildmann et al., 1992),

and in potato, protein phosphorylation is required for TD elicita-

tion by JA. TD is also highly expressed in flowers and has a

chloroplast transit peptide in the N-terminal region (Samach

et al., 1991, 1995). A strong association between JA signaling

and TD expression can be inferred from the synthesis of

JA–amino acid conjugates and suggests a mechanism linking

TD activity and herbivore resistance.

JA synthesis begins in plastids. There, a-linolenic acid is

oxygenated by lipoxygenase (LOX); converted to 12-oxo-phyto-

dienoic acid by allene oxide synthase and allene oxide cyclase

before being exported to the peroxisome; and reduced by

12-oxo-phytodienoic acid reductase. JA is produced after three

consecutive b-oxidation steps in the peroxisome (Li et al., 2005).

JA can be subsequently methylated to its volatile counterpart,

methyl jasmonate (MeJA), or conjugated with various sugars and

amino acids (Sembdner and Parthier, 1993; Sembdner et al.,

1994). An Arabidopsis thaliana gene (JASMONATE-RESISTANT1

[JAR1]) involved in JA responsiveness was shown to adenylate JA

before its conjugation with amino acids, of which the JA–isoleucine

conjugate (JA-Ile) was the most abundant (Staswick et al., 2002;

Staswick and Tiryaki, 2004). Because JA signaling is essential

for resistance to a large number of herbivore taxa (Halitschke

and Baldwin, 2003), TD might supply Ile for conjugation with JA

at the attack site and thereby function in defense signaling.

Another hypothesis for a defensive role for TD is based on a

recent proteomic analysis of the midgut contents of Manduca

sexta larvae that fed on tomato (Chen et al., 2005). This exciting

report revealed that one of the abundant proteins in the larval

midgut was TD, but TD that lacked a regulatory domain. This

truncated TD might efficiently degrade Thr without being inhib-

ited by Ile and function as an antinutritive defense by limiting the

supply of Thr needed for herbivore growth (Chen et al., 2005).

1 To whom correspondence should be addressed. E-mail [email protected]; fax 49-3641-571102.The author responsible for distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors (www.plantcell.org) is: Ian T. Baldwin([email protected]).W Online version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.106.041103

The Plant Cell, Vol. 18, 3303–3320, November 2006, www.plantcell.org ª 2006 American Society of Plant Biologists

Nicotiana attenuata is a particularly useful system in which to

study herbivore resistance responses. Not only is it well estab-

lished that JA signaling mediates herbivore resistance in the

field for this species (Baldwin, 1998; Kessler and Baldwin,

2001; Kessler et al., 2004), but also the direct and indirect de-

fense traits with which JA signaling influences herbivore resis-

tance are known (Halitschke et al., 2004; Steppuhn et al., 2004).

The responses of N. attenuata to one particular herbivore, the

solanaceous specialist M. sexta, are particularly well under-

stood. The attacked plant reorganizes its wound response when

eight fatty acid–amino acid conjugates, present in the herbivore’s

oral secretions (OS), are introduced into plant wounds during

feeding. The reorganization begins with a dramatic JA burst in the

attacked leaves (Schittko et al., 2000), which alters the expression

of numerous genes and the accumulation and release of second-

ary metabolites (Halitschke et al., 2000, 2001, 2003; Kahl et al.,

2000; Roda et al., 2004). Silencing the expression of the specific

lox that supplies the fatty acid hydroperoxides for JA biosynthesis

in N. attenuata (LOX3) reduces the OS-elicited JA burst and all

associated changes in the plant’s resistance traits (Halitschke and

Baldwin, 2003; Kessler et al., 2004).

Many of N. attenuata’s herbivore-responsive genes have been

identified by cDNA differential display, subtractive hybridization,

and cDNA-amplified fragment-length polymorphism display

(Halitschke et al., 2001, 2003; Hermsmeier et al., 2001; Schittko

et al., 2001; Hui et al., 2003; Voelckel and Baldwin, 2003). These

genes have been spotted onto microarrays, and their expression

behavior has been analyzed in response to various environmen-

tal stresses (Halitschke et al., 2003; Hui et al., 2003; Izaguirre

et al., 2003; Voelckel and Baldwin, 2003, 2004; Lou and Baldwin,

2004). In these experiments, TD expression was consistently

found to correlate with elicited herbivore resistance. TD was

cloned by differential display RT-PCR, found to be encoded by a

single gene, and strongly elicited when plants were attacked by

M. sexta larvae, mechanically wounded, or treated with MeJA;

neither Tobacco mosaic virus nor treatment with Agrobacterium

tumefaciens infection, ethylene, or methyl salicylate elicited TD

expression (Hermsmeier et al., 2001). Wounding and OS elicita-

tion increase TD expression, not only in the wounded leaf but

also in distal nonwounded leaves that are phyllotactically con-

nected by common orthostichies (Schittko et al., 2001). The

wound-induced expression of TD is reduced in N. attenuata

plants transformed with N. attenuata LOX3 in an antisense

orientation, demonstrating that TD elicitation requires JA signal-

ing (Halitschke and Baldwin, 2003). These observations suggest

that TD may be involved in defense against herbivore attack.

To examine the effect of TD on defense responses, we first

expressed 1.3 kb of the N. attenuata TD in an antisense orien-

tation. Transformed lines were readily characterized as having

one of two growth phenotypes: (1) plants with severely reduced

TD expression and activity, and stunted growth and develop-

ment (asTDS plants), and (2) plants with mildly reduced TD

expression and activity but otherwise wild-type growth and

development patterns (asTDM plants). Because plant–herbivore

interactions are difficult to interpret in plants that are severely

stunted in their growth and development, we also silenced TD

with a virus-induced gene-silencing (VIGS) system optimized for

N. attenuata (Saedler and Baldwin, 2004), which allowed us to

silence TD in wild-type plants. Finally, we cloned JAR4, the

Arabidopsis JAR1 homolog in N. attenuata, to demonstrate that

JAR4 conjugates Ile to JA to mediate defense signaling and

resistance to M. sexta larvae. We also tested the hypothesis that

TD functions as an antinutritive defense by adding Thr and Ile to

wild-type and TD-silenced plants and examined the conse-

quences of this supplementation for larval growth. The results of

this work support both hypotheses: TD plays an important role in

herbivore resistance by mediating JA-Ile signaling and also acts

as an antinutritional protein by depleting Thr levels.

RESULTS

We measured TD transcript accumulation in wild-type plants

after elicitation by insect attack and MeJA treatment. TD mRNA

in wild-type plants is strongly increased after attack from M.

sexta larvae (62-fold); TD transcripts are also strongly elicited

when leaves are wounded and treated with M. sexta OS (16-fold)

or when MeJA is applied in a lanolin paste to intact leaves

(79-fold) (see Supplemental Figure 1 online). Attack from other

leaf-chewing insect herbivores (Heliothis virescens and Spodop-

tera exigua) as well as from a species that feeds by lacerating and

flushing cells (Tupiocoris notatus) also strongly elicits TD tran-

script accumulation (18- to 41-fold; see Supplemental Figure

1 online), suggesting that TD is involved in plant defense.

To examine the function of TD, we first produced transgenic

plants expressing TD in an antisense orientation. T2 homozygous

plants from independently transformed lines, each harboring a

single copy of the transgene, as verified by segregation analysis

for antibiotic resistance and DNA gel blot analysis (see Supple-

mental Figure 2 online), were analyzed. Transformed lines were

readily characterized as having one of two growth phenotypes:

(1) plants with greatly reduced TD expression and activity and

retarded growth (asTDS plants; Figure 1A), and (2) plants with

mildly reduced TD expression and activity but whose growth

and development patterns were indistinguishable from those of

wild-type plants (asTDM plants; Figure 1A). Second, we pro-

duced TD-silenced plants using the VIGS method. TD-silenced

(TDVIGS) plants had less TD expression and activity but similar

growth patterns compared with empty vector (EV) control VIGS

plants (Figure 1A). All transgenic lines and VIGS plants were also

analyzed for levels of defense-related secondary metabolites

such as trypsin protease inhibitor (TPI) and nicotine.

Silencing TD Transcripts Decreases a-KB Accumulation

and Impairs Herbivore Resistance without Influencing

Plant Growth

To determine whether TD mRNA levels were suppressed in asTD

lines when plants’ leaves were treated with MeJA, TD mRNA levels

were analyzed by RNA gel blot. After MeJA treatment, levels of

TD mRNA in both asTDM and asTDS1 lines were reduced by

30 and 95% compared with wild-type levels. As expected,

antisense-oriented TD mRNA was found only in the asTD lines

(Figure 1B). Silencing the expression of TD transcripts translated

into changes in TD activity, which were assayed by measuring

a-KB, the product of TD. Before MeJA treatment, a-KB levels

were similar in wild-type and asTDM plants (Figure 1C; unpaired

3304 The Plant Cell

Figure 1. Suppressing TD in asTD and TDVIGS Plants.

(A) Wild-type (55-d-old), asTDM2 (55-d-old), asTDS1 (85-d-old), EV (54-d-old), and TDVIGS (54-d-old) plants. Note that asTDM2 plants are

morphologically indistinguishable from wild-type plants, but asTDS1 plants are severely stunted in their growth and morphologically different from wild-

type plants. TDVIGS and EV plants grow similarly.

(B) and (C) Accumulation of TD transcripts (B) and a-KB concentration (C) in a pooled sample of four replicate nodeþ1 leaves, which were treated with20 mL of lanolin containing 150 mg of MeJA and harvested after 24 h from wild-type and three independently transformed T2 asTD plants (asTDS1,

asTDM1, and asTDM2). The arrow indicates endogenous TD RNA (TD), and the arrowhead indicates antisense TD RNA (asTD). Ethidium bromide–

stained 18S rRNA was used as a loading control. Asterisks represent significant differences between MeJA-treated wild-type and MeJA-treated asTD

plants (unpaired t test: * P < 0.05; ** P < 0.01; **** P < 0.0001).

(D) to (G) Accumulation of TD transcripts ([D] and [F]) and a-KB concentration ([E] and [G]) in leaves at nodeþ1 from three replicate wild-type, asTDS1,and asTDM2 plants, which were wounded with a fabric pattern wheel and immediately treated with 20 mL of deionized water (W) or 20 mL of OS.

Asterisks represent significant differences between wild-type and asTD plants (two-way ANOVA, Fisher’s PLSD: **** P < 0.0001).

(H) and (I) Accumulation of TD transcripts (H) and a-KB (I) in TDVIGS plants. Plants were inoculated with Agrobacterium harboring tobacco rattle virus

(TRV) constructs that contain an EV or a 335-bp TD fragment (TDVIGS). Fourteen days after inoculation, leaves at node þ1 from four to five replicate EV

Threonine Deaminase in Defense Signaling 3305

t test, P # 0.546), but compared with wild-type plants, a-KB

levels in asTDS1 plants were reduced significantly (Figure 1C;

unpaired t test, P < 0.0001). Eightfold increases in levels of a-KB

were measured in wild-type plants 24 h after MeJA elicitation

(Figure 1C). When asTD plants were elicited, the levels of a-KB in

asTDM1, asTDM2, and asTDS1 plants were reduced signifi-

cantly—by 19, 33, and 80%, respectively—compared with those

in wild-type plants (Figure 1C; unpaired t test, P # 0.0498).

Plants treated with MeJA in a lanolin paste are continuously

elicited, as the MeJA slowly diffuses into the plant (Zhang et al.,

1997). To examine the effects of a more subtle elicitation treat-

ment, transgenic lines and VIGS plants were wounded with a

fabric pattern wheel and treated with water or M. sexta OS. TD

mRNA expression was analyzed by real-time PCR. When

wounded leaves were treated with water, TD mRNA attained

maximum values in wild-type plants 1.5 h after wounding and

waned slowly thereafter. Levels in both asTDM2 and asTDS1

plants were significantly lower than the levels in wild-type plants

(Figure 1D; Fisher’s protected least squares difference [PLSD],

P < 0.0001). The production of a-KB was slightly increased by

wounding. Although a-KB levels in asTDM2 plants were re-

duced, they did not differ significantly compared with the levels in

wild-type plants (Figure 1E; Fisher’s PLSD, P¼ 0.243), but levelsof a-KB in asTDS1 plants were reduced significantly compared

with the levels in wild-type plants (Figure 1E; Fisher’s PLSD, P <

0.0001). Leaves treated with water or OS showed similar ex-

pression patterns. TD mRNA levels in leaves treated with M.

sexta OS from both asTDM2 and asTDS1 plants were signifi-

cantly lower than the levels in leaves from wild-type plants

(Figure 1F; Fisher’s PLSD, P < 0.0001). Levels of a-KB in asTDM2

plants were reduced but did not differ significantly compared

with the levels in wild-type plants (Figure 1G; Fisher’s PLSD, P¼0.1969); however, levels of a-KB in asTDS1 plants were reduced

significantly compared with the levels in wild-type plants (Figure

1G; Fisher’s PLSD, P < 0.0001). TDVIGS plants also showed re-

duced TD mRNA levels compared with EV control plants. When

wounded leaves were treated with water or M. sexta OS and

then compared with EV plants, the levels of TD mRNA and a-KB

in TDVIGS plants were 80 and 71% lower in transcripts and

48 and 47% lower in TD activity (Figures 1H and 1I; unpaired

t test, P # 0.012).

To determine whether the mild suppression of TD transcripts

and activity observed in the asTDM lines influenced plant growth

and competitive ability, we synchronized the germination and

growth of the different lines and grew them individually in 2-liter

pots or in competition with each other in 2-liter pots. We mea-

sured stalk elongation, which previous experiments have re-

vealed to accurately measure competitive ability and relative

fitness (Glawe et al., 2003). No differences in stalk elongation

among the lines were observed when plants were grown singly or

in competition with wild-type plants (see Supplemental Figure 3

online). When TDVIGS and EV plants were grown individually in

1-liter pots, stalk lengths appeared not to differ (Figure 1A).

To determine whether TD is involved in plant defense, we

measured the performance of the insect herbivore M. sexta,

which is responsible for the largest losses in leaf area among

N. attenuata plants growing in nature (Baldwin, 1998). asTDS1

plants were severely stunted in their growth, and their leaf

developmental traits differed from those of wild-type plants

(Figure 1A), which confounded comparisons of herbivore per-

formance between wild-type and asTDS1 plants. Therefore, we

first compared herbivore performance on wild-type and the

morphologically indistinguishable asTDM lines. Freshly eclosed

M. sexta larvae placed on the source–sink transition leaf of each

of seven replicate plants of each genotype gained significantly

more mass on plants of both asTDM lines than they did on wild-

type plants. By day 6, larvae on asTDM2 plants had almost

doubled their mass compared with larvae on wild-type plants

(see Supplemental Figure 4A online; repeated-measurement

analysis of variance [ANOVA], F2,36 ¼ 15.988; P ¼ 0.0001;PLSD # 0.0485). Similarly, M. sexta larvae placed on VIGS

leaves gained significantly more mass on TDVIGS plants than

on EV plants (see Supplemental Figure 4B online; repeated-

measurement ANOVA, F1,21 ¼ 12.071; P ¼ 0.0023; PLSD ¼0.0023). These results demonstrate that reductions in TD ex-

pression and activity do not influence plant growth (even under

intense intraspecific competition [see Supplemental Figure 3

online]) but impair resistance to an adapted herbivore.

TD Silencing Impairs Elicited Direct Defenses in

asTD and TDVIGS Plants

To test the hypothesis that increasing Ile pools at the wound site

could be used for herbivore-elicited direct defenses, we mea-

sured TPI in TD-silenced plants. Compared with wounding alone,

OS treatment of puncture wounds in wild-type plants resulted in

a 2.4-fold increase in TPI activity (Figure 2A). The wound-induced

accumulation of TPI activity in asTDM2 plants was 31% lower

than that in wild-type plants (Figure 2A; unpaired t test, P ¼0.0911), and the OS-induced accumulation of TPI in asTDM2

plants was 41% lower than that in wild-type plants (Figure 2A;

unpaired t test, P ¼ 0.0386). The wound-induced accumulationof TPI activity in asTDS1 plants was 47% lower than that in wild-

type plants (Figure 2A; unpaired t test, P ¼ 0.0365), and the OS-induced accumulation of TPI in asTDS1 plants was 76% lower

than that in wild-type plants (Figure 2A; unpaired t test, P ¼0.002). However, OS treatment of wounds in asTDS1 plants did

not significantly increase TPI activity compared with water treat-

ment of wounds in asTDS1 plants (Figure 2A; unpaired t test, P¼0.719), suggesting that severe nutritional deficiencies inhibited

Figure 1. (continued).

and TDVIGS plants were wounded with a fabric pattern wheel and immediately treated with 20 mL of deionized water (W) or 20 mL of OS. Asterisks

represent significant differences between members of a pair (unpaired t test: * P < 0.05; ** P < 0.01).

The transcripts were analyzed by real-time PCR as means 6 SE of three to five replicate leaves in arbitrary units from a calibration with 53 dilution series of

cDNAs prepared from asTD plant RNA samples extracted 1 h after wounding ([D] and [F]) or from EV plant RNA samples extracted 1 h after wounding (H).

3306 The Plant Cell

TPI production in these plants. VIGS plants also showed induced

TPI activity when leaves were wounded and treated with water or

M. sexta OS (Figure 2B); however, when leaves were treated with

M. sexta OS, TPI activity in TDVIGS plants was reduced signif-

icantly compared with that in EV plants (Figure 2B; unpaired

t test, P ¼ 0.0006). These results suggest that diminished TPIlevels are a major cause of the increased performance of M.

sexta larvae feeding on asTDM2 and TDVIGS plants. When Ile

was added to water or OS before being applied to the puncture

wounds, TPI levels in both asTDM2 and TDVIGS plants were

restored to levels found in wild-type and EV plants. In asTDS1

plants, adding Ile to water restored TPI levels to those found in

wild-type plants (Figure 2A; unpaired t test, P¼ 0.4507). When Ilewas added directly to OS, TPI levels in asTDS1 plants were still

lower than those in wild-type plants (Figure 2A; unpaired t test,

P ¼ 0.0427). However, adding Ile to OS and then applying theseto the puncture wounds significantly increased TPI activity in

asTDS1 plants compared with OS-treated asTDS1 plants (Figure

2A; unpaired t test, P¼ 0.0159). The restoration of TPI activity byIle supplementation at the wound site in asTDM and TDVIGS

plants could be attributed either to the restoration of the biosyn-

thetic needs of TPI production or to its signaling.

Recent research on N. attenuata has demonstrated that

silencing the LOX required for JA biosynthesis also silences

inducible nicotine and TPI defenses and increases M. sexta larval

performance (Halitschke and Baldwin, 2003). Moreover, JA is

known to be conjugated to several amino acids in vitro, and

JA-Ile is the most abundant JA–amino acid conjugate in Arabi-

dopsis seedlings (Staswick and Tiryaki, 2004). Because TD is

involved in Ile synthesis, we examined whether the effect of

silencing TD on herbivore performance could be attributed to JA

signaling via JA-Ile synthesis or turnover.

TD Silencing Impairs JA Signaling in asTD

and TDVIGS Plants

More JA is elicited from wounded leaves treated with M. sexta

OS than from leaves that have only been wounded (Halitschke

et al., 2001). To determine whether M. sexta OS elicit the same

rapidly increasing and declining JA-Ile pools, leaves at node

þ1 from four independently treated plants from each genotype,wild type and asTDM2, were wounded, treated with OS (Figure

3A), and analyzed by LC-MS at each harvest time. As expected,

in treated wild-type leaves, a JA burst was elicited within 30 min,

reached maximum levels at 1 h, and declined rapidly after 1.5 h

(Figure 3B). Similar responses were observed in JA-Ile pools in

treated wild-type leaves (Figure 3C). The JA burst in asTDM2

plants was similar to that in wild-type plants but waned more

slowly at 1.5 h after elicitation (Figure 3B; unpaired t test, P ¼0.0120). The OS-elicited JA-Ile burst in asTDM2 plants was less

than that in wild-type plants, with pools being significantly lower

(36 and 68%) at 0.5 and 3 h, respectively (Figure 3C; unpaired

t test, P # 0.0237). To determine whether JA-Ile is synthesized

from JA and Ile at the wound site, nodeþ1 leaves from wild-typeplants were wounded and immediately treated with OS contain-

ing 0.625 mmol of [13C4]Thr or [13C6]Ile (Figure 3A). Four replicate

plants were harvested for each treatment and harvest time to

measure the elicited kinetics of JA and 13C-labeled JA-Ile by

LC-MS analysis. Adding [13C4]Thr and [13C6]Ile to OS reduced

the levels of JA compared with OS (Figures 3B and 3D). Adding

Thr to an OS-elicited wound reduced the maximum JA values by

;5.5 nmol/g fresh weight (cf. [13C4]Thr treatments: 6 nmol/gfresh weight in Figure 3D with OS-elicited values of 11.5 nmol/g

fresh weight in Figure 3B). Adding the more efficiently incorpo-

rated amino acid, Ile, reduced JA values even further (to 4 nmol/g

fresh weight; Figure 3D). The reduced levels of JA were used to

synthesize JA-Ile. Significant quantities of 13C-labeled JA-Ile

were detected when either [13C4]Thr or [13C6]Ile was applied,

demonstrating that [13C4]Thr was rapidly converted to Ile at the

wound site and used to synthesize 13C-labeled JA-Ile. As ex-

pected, Thr was incorporated less efficiently into JA-Ile than was

Ile (Figure 3E), demonstrating that the conjugation capacity of an

elicited leaf is limited by substrate availability. Compared with

wild-type plants, asTDM2 plants were less efficient in incorpo-

rating [13C4]Thr into 13C-labeled JA-Ile at 0.5 h after elicitation

(Figure 3F; unpaired t test, P ¼ 0.0183), suggesting that mildlysilencing TD expression correlated with detectable reductions in

the conversion of Thr to Ile and its subsequent incorporation into

JA-Ile. When wounds were treated with 0.625 mmol of JA (Figure

3A), plants sustained increased JA-Ile pools for 2.5 h (Figure 3G),

Figure 2. Silencing TD in asTD and TDVIGS Plants Impairs OS-Elicited

TPI Activity.

(A) Mean TPI activity (6SE) in wild-type, asTDM2, and asTDS1 plants of

three replicate node þ1 leaves that were harvested 3 d after beingwounded and treated with 20 mL of either deionized water (W) or M. sexta

OS supplemented with 0.625 mmol of Ile (þIle). Leaves from controlplants (C) were left intact and untreated. Asterisks represent significant

differences between members of a pair (unpaired t test: * P < 0.05;

** P < 0.01).

(B) Mean TPI activity (6SE) in EV and TDVIGS plants of four to five

replicate node þ1 leaves that were harvested 3 d after being woundedand treated with 20 mL of either deionized water (W) or M. sexta OS

supplemented with 0.625 mmol of Ile (þIle). Leaves from control plants(C) were left intact and untreated. Asterisks represent significant differ-

ences between members of a pair (unpaired t test: * P < 0.05; *** P < 0.01).


Figure 3. OS-Elicited JA and JA-Ile Are Regulated by Thr, Ile, and JA; Silencing TD in asTDM2 Plants Reduces JA-Ile.

(A) Numbering of leaf positions in rosette-stage N. attenuata plants. The leaf undergoing the source–sink transition (T) was designated as growing at

node 0. The treated leaf growing at node þ1, which is older by one leaf position than the source–sink transition leaf, was wounded with a fabric patternwheel, and the resulting puncture wounds (W) were immediately treated with 20 mL of M. sexta OS, OS containing 0.625 mmol of 13C4-labeled Thr (13C4Thr) or 13C6-labeled Ile (13C6 Ile), water containing 0.625 mmol of JA (JA), or water containing 0.625 mmol of JA and 13C6-labeled Ile. The treated leaves

were harvested to measure JA, JA-Ile, and isotope-labeled JA-Ile.

(B) and (C) Mean 6 SE JA (B) and JA-Ile (C) concentrations in OS-treated leaves of four replicate wild-type and asTDM2 plants. Asterisks represent

significant differences between members of a pair analyzed at the same time after OS elicitation (unpaired t test: * P < 0.05). FW, fresh weight.

(D) and (E) Mean 6 SE JA (D) and isotope-labeled JA-Ile (E) concentrations in [13C4]Thr- or [13C6]Ile-treated leaves of three replicate wild-type plants.

Asterisks represent significant differences between members of a pair (unpaired t test: * P < 0.05; *** P < 0.001; **** P < 0.0001).

(F) Mean 6 SE isotope-labeled JA-Ile concentrations in [13C4]Thr-treated leaves of three replicate wild-type and asTDM2 plants. Asterisks represent

significant differences between members of a pair (unpaired t test: * P < 0.05).

(G) Mean 6 SE JA-Ile concentrations in JA-treated leaves of three replicate wild-type and asTDM2 plants. Asterisks represent significant differences

between wild-type and asTDM2 plants (two-way ANOVA, Fisher’s PLSD: ** P < 0.01).

(H) Mean 6 SE isotope-labeled JA-Ile concentrations in JA- and [13C6]Ile-treated leaves of three replicate wild-type and asTDM2 plants.

3308 The Plant Cell

demonstrating that the level of JA regulates the level of JA-Ile.

Under these experimental conditions, when JA is not limited,

clear differences between the abilities of asTDM2 plants and

wild-type plants to produce JA-Ile were readily discerned: the

amount of JA-Ile in asTDM2 plants was significantly lower than

that in wild-type plants (Figure 3G; two-way ANOVA, F1,16 ¼9.768; P ¼ 0.0065). When both [13C6]Ile and JA were applied towounded wild-type and asTDM2 plants, the levels of JA-Ile did

not differ (Figure 3H; two-way ANOVA, F1,16 ¼ 0.999; P ¼0.3324). These results demonstrate that Ile limits JA-Ile synthesis

in asTDM2 plants more than in wild-type plants and that the

JA-Ile conjugation enzyme in asTDM2 plants is equally active in

wild-type plants. These experiments also demonstrate that the

JA and JA-Ile bursts that erupt when M. sexta OS are introduced

into a wound can be simulated by adding JA to a wound.

To determine the effect of silencing TD on JA and JA-Ile

elicitation in asTDS1 and TDVIGS plants, leaves were wounded

and treated with OS or JA. Four to five independently treated

plants from each genotype were analyzed at each harvest time

(Figure 4A). The OS-elicited changes in JA and JA-Ile pools in

asTDS1 plants did not resemble the bursts observed in either

wild-type or asTDM plants. Both JA and JA-Ile pools waxed and

waned slowly, attaining maximum values at 2 h (Figures 4B and

4C). The integrated JA levels in asTDS1 plants (;33.66 nmol/gfresh weight per 5 h; Figure 4B) were 18% higher than those in

wild-type plants (;28.53 nmol/g fresh weight per 5 h; Figure 4B).The integrated JA-Ile levels in asTDS1 plants (;2.61 nmol/gfresh weight per 5 h; Figure 4C) were 25% lower than those in

wild-type plants (;3.46 nmol/g fresh weight per 5 h; Figure 4C).The incorporation of [13C4]Thr into 13C-labeled JA-Ile in asTDS1

Figure 4. Silencing TD in asTDS1 and TDVIGS Plants Reduces JA-Ile.

(A) Nodeþ1 leaves were wounded with a fabric pattern wheel and the resulting puncture wounds (W) immediately treated with 20 mL of M. sexta OS, OScontaining 0.625 mmol of 13C4-labeled Thr (13C4 Thr), or water containing 0.625 mmol of JA (JA). The treated leaves were harvested to measure JA,

JA-Ile, and isotope-labeled JA-Ile.

(B) and (C) Mean 6 SE JA (B) and JA-Ile (C) concentrations in OS-treated leaves of four replicate wild-type and asTDS1 plants. Asterisks represent

significant differences between members of a pair (unpaired t test: * P < 0.05; ** P < 0.01; *** P < 0.001). FW, fresh weight.

(D) Mean 6 SE isotope-labeled JA-Ile concentrations in [13C4]Thr-treated leaves of three replicate wild-type and asTDS1 plants. Asterisks represent

significant differences between members of a pair (unpaired t test: ** P < 0.01; *** P < 0.001).

(E) and (F) Mean 6 SE JA-Ile concentrations in leaves of five replicate EV and TDVIGS plants. Leaves were harvested 2 h after JA (E) or 1 h after OS (F)

treatment. Asterisks represent significant differences between EV and TDVIGS plants (unpaired t test: * P < 0.05; ** P < 0.01).

(G) Mean 6 SE isotope-labeled JA-Ile concentrations of five replicate EV and TDVIGS plants 1 h after leaves were wounded and treated with OS and

[13C4]Thr. Asterisks represent significant differences between EV and TDVIGS plants (unpaired t test: ** P < 0.01).


plants was detected only 0.5 h after elicitation, and 13C-labeled

JA-Ile levels were 25% of those in wild-type plants (Figure 4D;

unpaired t test, P # 0.0036). The lower Ile pools of asTD plants

may account for the slower decline of the JA burst and for the

lower levels of JA-Ile observed in these plants. Compared with

EV plants, TDVIGS plants also showed reduced levels of JA-Ile.

When wounds were treated with JA or OS, JA-Ile levels in

TDVIGS plants at 1 h after elicitation were 24 and 30% of those in

EV plants (Figures 4E and 4F; unpaired t test, P # 0.043). The

levels of [13C4]Thr incorporated into 13C-labeled JA-Ile in TDVIGS

plants were 26% of those in EV plants (Figure 4G; unpaired t test,

P ¼ 0.005), demonstrating that JA-Ile synthesis is limited by theIle produced by TD at the wound site.

Supplementing asTD Plants with JA-Ile Restores Direct

Defenses and Herbivore Resistance

After discovering that asTDM2 plants had reduced levels of

JA-Ile when plant wounds were treated with JA (Figure 3G), we

wanted to determine whether JA-Ile could elicit direct defenses

and whether the herbivore resistance of asTDM2 plants could be

restored by JA-Ile treatment. The compounds were added to

wounds in aqueous solutions, because these water-soluble com-

pounds are unable to transverse the leaf cuticle when applied in a

lanolin paste.

Adding JA or JA-Ile to wounds of both wild-type and asTDM2

plants significantly increased TPI above the levels reached when

nothing was added to wounds (Figure 5A; unpaired t test, P #

0.0035). JA addition, which was demonstrated to produce sus-

tained differences in endogenous JA-Ile levels between wild-

type and asTDM2 plants (Figure 3G), elicited TPI in asTDM2

plants at levels that were 40% lower than those in wild-type

plants (Figure 5A; unpaired t test, P¼ 0.0361). When plants weretreated with JA-Ile, the induced TPI responses did not differ

between wild-type and asTDM2 lines (Figure 5A; unpaired t test,

P ¼ 0.8197), although they were significantly lower than the TPIresponses elicited in wild-type plants by JA treatment. Higher

levels of nicotine resulted when plants were treated with JA or

JA-Ile and not only wounded (Figure 5A; unpaired t test, P #

0.0022). Nineteen percent less nicotine accumulated in response

to JA treatment in asTDM2 plants compared with wild-type

plants (Figure 5A; unpaired t test, P ¼ 0.0396), whereas theresponses to JA-Ile treatment did not differ between asTDM2

and wild-type plants (Figure 5A; unpaired t test, P ¼ 0.4214).However, unlike levels of TPI, the levels of nicotine elicited in

plants treated with JA-Ile were much higher than in plants treated

with JA (Figure 5A). Levels of chlorogenic acid in JA or JA-Ile

treatment were the same in the untreated control. Although

similar levels of diterpene glycosides were elicited by either JA or

JA-Ile treatment, the levels did not differ between asTDM2 and

wild-type plants (see Supplemental Figure 5 online), demon-

strating that these secondary metabolites are not differentially

Figure 5. JA-Elicited Herbivore Resistance, Nicotine, and TPI Produc-

tion Are Impaired in asTDM2 Plants Compared with Wild-Type Plants but

Restored by Adding JA-Ile.

(A) Mean (6SE) TPI and nicotine levels in leaves from three replicate wild-

type and asTDM2 plants growing at node þ1, 3 d after being woundedand treated with 20 mL of deionized water (W), Ile (WþIle), JA (WþJA), orJA-Ile conjugate (WþJA-Ile), all at 0.625 mmol. Asterisks representsignificant differences between members of a pair (unpaired t test:

* P < 0.05). FW, fresh weight.

(B) Mean (6SE) mass of M. sexta larvae after 3, 6, and 9 d of feeding on 16

replicate wild-type plants and two lines of T2 transgenic plants (asTDM2

and asLOX3). Leaves were treated with 0.625 mmol of JA-Ile or left

untreated (C). Top graph, asterisks represent significant differences

between untreated wild-type plants and two lines of untreated T2

transgenic plants on day 9 (unpaired t test: * P < 0.05; *** P < 0.001).

Bottom graph, asterisks represent significant differences between un-

treated and JA-Ile–treated plants on day 9 (unpaired t test: ** P < 0.01).

asLOX3 plants, which are largely defenseless because of their impaired

JA signaling (Halitschke and Baldwin, 2003), were included as a positive

control for herbivore resistance.

3310 The Plant Cell

elicited by JA and JA-Ile. Differences in the ability of JA and JA-Ile

to elicit nicotine and TPI may reflect different rates of absorption

in treated leaves or their transport within the plant. Most impor-

tant, the elicited nicotine and TPI responses, which were signif-

icantly lower in JA-treated asTDM2 plants than in wild-type

plants, did not differ between wild-type and asTDM2 plants when

plants were treated with JA-Ile. These results demonstrated that

JA-Ile could restore the direct defense responses of asTDM2

plants to the levels of these responses in wild-type plants; the

next step was to determine whether resistance to M. sexta larvae

could be similarly restored.

We measured the performance of the M. sexta larvae on wild-

type and asTDM2 plants and a genotype of N. attenuata plants

(asLOX plants) in which LOX3, the lipoxygenase gene supplying

fatty acid hydroperoxides for JA biosynthesis, was silenced by

antisense expression. asLOX plants have lower levels of JA and

reduced levels of the direct defenses, nicotine and TPIs, and

therefore are impaired in their herbivore resistance (Halitschke

and Baldwin, 2003). These defenseless plants were included in

the analysis to gauge the degree to which herbivore resistance

had been impaired in the asTDM lines. By day 9, M. sexta larvae

that fed on untreated asTDM2 and asLOX plants had gained 68

and 166% more mass than those that fed on wild-type plants,

respectively (Figure 5B; unpaired t test, P # 0.041). Treating

asTDM plants with JA-Ile fully restored the plant’s resistance;

larvae that fed on JA-Ile–treated asTDM plants attained masses

that were statistically indistinguishable from those that fed on

JA-Ile–elicited wild-type plants (Figure 5B; unpaired t test, P ¼0.19). These results demonstrate that JA-Ile is a potent elicitor of

direct defenses, particularly TPI and nicotine, and that treatment

of asTDM2 plants with JA-Ile can restore this line’s resistance to

M. sexta larvae.

Because we now understood that the traits responsible for the

defects in herbivore resistance were associated with TD silenc-

ing, we were ready to examine herbivore resistance in the

developmentally challenged asTDS plants. M. sexta larvae that

fed on JA-treated asTDS plants gained less mass compared with

those that fed on untreated asTDS plants, but the difference was

not significant by day 9 (Figure 5B; unpaired t test, P ¼ 0.134).Feeding on JA-Ile–treated asTDS plants, however, the larvae

gained significantly less mass (45%) compared with those that

fed on untreated asTDS plants by day 9 (Figure 5B; unpaired

t test, P¼ 0.0027). These results demonstrate that even in plantswith severely silenced TD that suffer from severe nutritional

deficiencies, Ile is conjugated to JA at the wound site to mediate

defense signaling. Supplementing wounds with JA-Ile restores a

modicum of induced resistance in these severely growth-

impaired plants.

Suppressing TD and JAR4 by VIGS Impairs JA Signaling

and Herbivore Resistance

To further examine whether JA-Ile is the signal molecule that

elicits herbivore resistance, we cloned the Arabidopsis JAR1

homolog JAR4 (GenBank accession number DQ359729) from N.

attenuata using RT-PCR. To investigate whether JAR4 encodes

the enzyme that conjugates amino acids to JA in N. attenuata, we

collected amino acid sequences of JAR-like proteins using N.

attenuata JAR4 as a query. Phylogenetic analysis revealed that

these proteins clustered into three groups; JAR4 and JAR1

cluster together with three functionally unknown proteins (see

Supplemental Figure 6 online) that share >60% amino acid

identity (see Supplemental Figure 7 online), suggesting that they

share similar functions as JAR1, namely, conjugating amino acid

to JA (Staswick et al., 2002; Staswick and Tiryaki, 2004). The

other Arabidopsis JAR family members, GH3.1, GH3.2, GH3.5,

and GH3.17, which conjugate amino acids to indole-3-acetic

acid (Staswick et al., 2005), clustered together in a separate

group. DNA gel blotting revealed that JAR4 is a single-copy gene

in the N. attenuata genome (see Supplemental Figure 8 online).

These results suggested that JAR4 is a good candidate for the

JA-conjugating enzyme in N. attenuata.

To determine whether JAR4 mRNA is elicited by wounding or

OS treatment of wounds, plants were wounded with a fabric

pattern wheel and treated with water or OS, and JAR4 mRNA

accumulation was analyzed by quantitative real-time PCR. In

response to wounding alone, JAR4 mRNA levels increased within

30 min, reached a maximum at 1.5 h, and declined after 3 h (Figure

6). Similar patterns of transcript accumulation were observed in

OS-treated wild-type leaves, but these levels waned more slowly

and did not return to control levels after 12 h (Figure 6).

To determine whether JAR4, like TD, is involved in eliciting

herbivore resistance, we used the VIGS system optimized for N.

attenuata (Saedler and Baldwin, 2004) to silence JAR4 and TD

mRNA separately in wild-type plants. TD RNA levels in TDVIGS

plants were 20, 17, or 16% of those in EV control plants when

plants were untreated, attacked by M. sexta larvae, or treated

with JA-Ile, respectively (Figure 7A; unpaired t test, P # 0.035).

JAR4 RNA levels in JAR4VIGS plants were 27, 38, or 49% of

those in EV plants when plants were untreated, attacked by

M. sexta larvae, or treated with JA-Ile, respectively (Figure 7A;

unpaired t test, P # 0.032). Analyzing JA-Ile pools 1 h after OS

elicitation in the VIGS plants demonstrated that both TD and

JAR4 are important in JA-Ile synthesis; elicited JA-Ile levels in

TDVIGS and JAR4VIGS plants were 30 and 29% of those in EV

plants (Figure 7B; unpaired t test, P # 0.042). Adding JA to the

Figure 6. Accumulation of JAR4 Transcripts after Elicitation by Wound-

ing and OS Treatments.

Leaves at node þ1 were wounded with a fabric pattern wheel, and theresulting wounds were immediately treated with 20 mL of deionized water

(W) or with M. sexta OS in five replicate wild-type plants. The transcripts

were analyzed by real-time PCR as means 6 SE of five replicate leaves in

arbitrary units from calibration with a 53 dilution series of cDNAs

prepared from RNA samples extracted at 1 h after wounding.


wound sites of either elicited TDVIGS or JAR4VIGS plants could

not restore the JA-Ile accumulation observed in EV plants (Figure

7B; unpaired t test, P # 0.0062).

As was demonstrated for asTD transgenic plants compared

with EV plants, TDVIGS and JAR4VIGS plants were both highly

susceptible to attack by M. sexta larvae. When M. sexta larvae

were placed on untreated leaves, larvae gained significantly

more mass on both TDVIGS and JAR4VIGS plants than they did

on EV plants. By day 6, their masses were already twice those of

larvae on EV plants. By day 9, larvae that fed on TDVIGS and

JAR4VIGS plants had gained 80% more weight than those that

fed on EV plants (Figure 8A, control; repeated-measurement

ANOVA, F2,31¼ 4.634; P¼ 0.017; PLSD # 0.037). When M. sextalarvae were placed on JA-Ile–treated leaves and weighed on

days 6 and 9, larvae that fed on TDVIGS and JAR4VIGS plants

had attained masses that were statistically indistinguishable

from those that fed on EV plants (Figure 8A, JA-Ile; repeated-

measurement ANOVA, F2,30¼ 2.473; P¼ 0.010; PLSD $ 0.147).Like wild-type plants, VIGS plants also showed increased TPI

levels when attacked by M. sexta larvae or treated with JA-Ile

(Figure 8B). When plants were attacked by M. sexta larvae,

elicited TPI levels in TDVIGS and JAR4VIGS plants were 22 and

35% of those in EV plants (Figure 8B; unpaired t test, P # 0.023),

demonstrating that both TD and JAR4 are involved in TPI

elicitation. When plants were treated with JA-Ile, the induced

TPI levels in TDVIGS and JAR4VIGS plants were restored to

those of EV plants (Figure 8B; unpaired t test, P $ 0.627),

demonstrating that TPI activity was not affected by VIGS inoc-

ulation and that treatment of TD- and JAR4-silenced plants with

JA-Ile restored elicited TPI activity and herbivore resistance in

TDVIGS and JAR4VIGS plants. These results demonstrate that

the decrease in herbivore resistance in TD- or JAR4-silenced

plants could be attributed to decreases in defense responses

associated with inhibited JA-Ile signaling. However, recently it

was suggested that TD could function as an antinutritive defense

by depleting Thr in the herbivore midgut (Chen et al., 2005);

therefore, we also examined whether the effect of TD on herbi-

vore performance could be attributed to amino acid depletion by

supplementing TD-silenced and wild-type plants with Thr and Ile

and measuring TD activity in larval frass.

Thr Supplementation of TD-Silenced Plants Increases

Herbivore Performance, whereas Ile Supplementation

Restores Herbivore-Resistance Traits

To test the hypothesis that herbivore-elicited TD functions as an

antinutritive defense by depleting Thr levels in the M. sexta

midgut, we treated EV and TDVIGS plants daily with either water

or 0.25 M Thr or Ile and allowed larvae to feed on these plants.

One hour after elicitation with OS, JA-Ile levels in water-treated

TDVIGS plants were significantly lower (64%) than those in

water-treated EV plants (Figure 9A; unpaired t test, P ¼ 0.0068).JA-Ile levels in Thr-treated TDVIGS plants were also lower (55%)

than those in EV plants (Figure 9A; unpaired t test, P ¼ 0.0028);however, adding Ile to OS restored the JA-Ile levels of TDVIGS

plants to those of EV plants (Figure 9A; unpaired t test, P ¼0.0809). The reduced levels of JA-Ile were reflected in TPI

production. Elicited TPI levels in TDVIGS plants were 55 and

Figure 7. Silencing TD and JAR4 by VIGS Reduces Transcript and JA-Ile

Accumulation.

(A) VIGS of TD and JAR4 transcripts. Plants were inoculated with

Agrobacterium harboring TRV constructs, which contain an EV, a 335-

bp TD fragment (TDVIGS), or a 292-bp JAR4 fragment (JAR4VIGS).

Fourteen days after inoculation, leaves were wounded, treated with

0.625 mmol of JA-Ile, and harvested 1 h later; or they were made

available for M. sexta larvae to feed on for another 12 d, after which they

were harvested (H); or they were harvested immediately from untreated

plants (C). The transcripts were analyzed by real-time PCR as means 6

SE of five replicate leaves in arbitrary units from a calibration with a 53

dilution series of cDNAs prepared from EV control RNA samples.

Asterisks represent significant differences between members of a pair

(unpaired t test: * P < 0.05; ** P < 0.01; **** P < 0.0001).

(B) Mean 6 SE JA-Ile concentrations in leaves of four to five replicate EV,

TDVIGS, and JAR4VIGS plants. Fourteen days after inoculation, leaves

were wounded, treated with 20 mL of either M. sexta OS or water

containing 0.625 mmol of JA, and harvested 1 h later. Asterisks represent

significant differences between EV and VIGS plants (unpaired t test: * P <

0.05; ** P < 0.01). FW, fresh weight.

3312 The Plant Cell

47% lower than those in EV plants when plants were treated with

water or Thr during M. sexta larval feeding (Figure 9B; unpaired

t test, P # 0.004). Treatment with Ile restored induced TPI levels

in TDVIGS plants to those in EV plants (Figure 9B; unpaired t test,

P ¼ 0.2776). However, M. sexta larvae that fed on Thr- or Ile-supplemented EV plants gained significantly more mass (48 and

85%) than those that fed on water-treated EV plants (Figure 9C;

unpaired t test, P # 0.0127). Similarly, M. sexta larvae that fed on

Thr-supplemented TDVIGS plants gained more mass (57%) than

did those that fed on water-treated TDVIGS plants (Figure 9C;

unpaired t test, P ¼ 0.048). By contrast, larvae that fed on Ile-supplemented TDVIGS plants did not differ from those that fed

on water-treated TDVIGS plants (Figure 9C; unpaired t test, P ¼0.95). Interestingly, larvae that fed on Ile-supplemented TDVIGS

plants did not differ from those that fed on Ile-supplemented EV

plants (Figure 9C; unpaired t test, P ¼ 0.508). In summary,supplementing leaves with Thr, but not Ile, significantly increased

larval performance in TD-silenced plants, consistent with pre-

dictions that TD was functioning as a postingestive antinutritive

defense.

Tomato TD is active not only in the midgut but also in the frass of

feeding M. sexta larvae, and TD activity in midgut and frass is

negatively correlated with insect performance (Chen et al., 2005).

To evaluate the role of N. attenuata TD in M. sexta larvae, we first

measured TD activity in frass of M. sexta larvae that fed on either

N. attenuata or tomato plants. TD activity in frass of larvae that fed

on tomato plants was 14.3 6 1.1 mmol�min�1�g�1 dry mass, andTD activity in larvae that fed on N. attenuata was 12.3 6 0.7mmol�min�1�g�1 dry mass, demonstrating that tomato and N.attenuata TDs are similarly active in the frass of M. sexta larvae.

Levels of TD in frass of larvae that fed on Thr-supplemented EV

plants were similar to those in frass of larvae that fed on water-

treated EV plants (Figure 9D; unpaired t test, P¼ 0.476); however,levels of TD in frass of larvae that fed on Ile-supplemented EV

plants were 58% of those in frass of larvae that fed on water-

treated EV plants (Figure 9D; unpaired t test, P¼ 0.016), consistentwith the expectations of feedback inhibition to TD by Ile. Levels of

TD in frass of larvae that fed on TDVIGS plants were low and did

not differ among treatments (Figure 9D; unpaired t test, P $ 0.215).

Levels of TD in frass of larvae that fed on the water- or Thr-

supplemented TDVIGS plants were 54% of those of EV plants

(Figure 9D; unpaired t test, P # 0.041). Levels of TD in frass of

larvae that fed on Ile-supplemented TDVIGS and EV plants were

similar (Figure 9D; unpaired t test, P ¼ 0.214). These experimentsdemonstrated that herbivores that fed on EV plants realize small

benefits in growth performance from Thr and Ile supplementations

to their diet. However, in TDVIGS plants, herbivores benefit from

Thr but not from Ile supplementation. The strong, positive effect of

Thr on herbivore performance in TDVIGS plants implies that Thr

limits M. sexta growth and development. The negative effect of Ile

on herbivore performance in TDVIGS plants is consistent with a

role for Ile in defense activation via JA-Ile–mediated signaling.

DISCUSSION

Because of the discovery, more than two decades ago, of the

genes responsible for the biosynthesis of amino acids, plant

biologists were able to determine which were essential for

growth and development. In an attempt to improve the nutritional

value of cereal crops, which have low levels of Lys and Thr,

biologists have focused attention on the essential amino acids,

Thr, Lys, Met, and Ile, which are synthesized via a common

pathway (Azevedo et al., 1997). TD catalyzes the conversion of

Figure 8. Silencing TD and JAR4 by VIGS Reduces Herbivore Resis-

tance and TPI Activity; Adding JA-Ile Restores Them.

(A) Mean 6 SE mass of M. sexta larvae after 6 and 9 d of feeding on 10 to

13 replicate plants, each inoculated with Agrobacterium harboring TRV

constructs, which contain an EV, a 335-bp TD fragment (TDVIGS), or a

292-bp JAR4 fragment (JAR4VIGS). Fourteen days after inoculation,

leaves were either wounded and treated with 0.625 mmol of JA-Ile (JA-Ile)

or left untreated (control). Asterisks represent significant differences

between EV and VIGS plants (repeated-measurement ANOVA, Fisher’s

PLSD: * P < 0.05; ** P < 0.01).

(B) Mean 6 SE TPI activity of five replicate EV, TDVIGS, and JAR4VIGS

plants. Fourteen days after inoculation, leaves were wounded, treated

with 0.625 mmol of JA-Ile, and harvested 3 d later; or they were made

available for M. sexta larvae to feed on for another 12 d, after which they

were harvested (H); or they were harvested immediately from untreated

plants (C). Asterisks represent significant differences between members

of a pair (unpaired t test: * P < 0.05; ** P < 0.01).


Thr to a-KB, the first committed step in Ile biosynthesis

(Umbarger, 1978).

The research presented here highlights the challenges of

disentangling the multiple roles that the enzymes involved in

amino acid biosynthesis can play in plants as well as after the

plant has been ingested by an herbivore. This research also

highlights the value of analyzing subtle phenotypes in plants for

which the determinants of ecological performance are well

understood.

TD’s role in herbivore resistance was discovered with the

transformants (asTDM) in which TD expression was mildly si-

lenced (Figure 1). These plants had completely normal growth

phenotypes, even under stringent competition regimes, but their

resistance to herbivores was impaired (see Supplemental Figure

4 online), allowing researchers to understand TD’s unusual

transcriptional behavior in response to wounding, herbivore

attack, and JA elicitation (Hildmann et al., 1992; Halitschke

et al., 2001; Hermsmeier et al., 2001; Schittko et al., 2001). The

susceptibility of asTDM plants to attack from M. sexta larvae was

associated with the reduced levels of two inducible direct de-

fenses: TPIs and nicotine (Figures 2 and 5). Previous research

has demonstrated that silencing either of these defenses in

N. attenuata plants increases the susceptibility of plants to attack

from M. sexta larvae and enhances larval performance (Steppuhn

et al., 2004; Zavala et al., 2004a, 2004b). Moreover, both of these

direct defenses are elicited by JA signaling (Halitschke et al.,

2004). The kinetics of the JA and JA-Ile bursts induced by larval

elicitors were found to be subtly altered in asTDM plants (Figure

3). This observation led to the discoveries that JA is conjugated

with Ile at the wound site and that herbivory-elicited TD supplies

the Ile required for the formation of JA-Ile. Supplementing

wounds in asTDM plants with Ile restored the wild-type kinetics

of the JA-Ile burst and also elicited direct defenses. Treating

asTDM plants with JA-Ile restored the plant’s ability to elicit

direct defenses and thereby the resistance of asTDM plants to

attack from M. sexta larvae (Figure 5). These results highlight the

dynamic role that JA-Ile plays in defense signaling and suggest

that subtle changes in the kinetics of JA and JA-Ile accumulation

after herbivore attack can profoundly affect defense elicitation.

JA-Ile’s role as a defense signal was confirmed in the analysis

of the asTDS plants, in which all of the subtle changes in defense

signaling observed in asTDM plants were exaggerated. In asTDS

plants, the OS-elicited JA and JA-Ile bursts observed in wild-type

plants were much slower (Figure 4). The OS-elicited JA-Ile

production was lower (Figure 4), and JA-Ile treatment effectively

restored herbivore resistance (Figure 5). Hence, although the

developmental defects of asTDS plants prevented direct com-

parisons of herbivore resistance with that in wild-type plants,

some of the defensive deficiencies of asTDS plants could be

compensated for by Ile or JA-Ile treatment. The ability to com-

plement these defensive deficiencies in plants suffering from

severe nutritional deficiencies underscores the importance of

Figure 9. Effects on Herbivore Resistance of Thr or Ile Supplementation

to Leaves of TD-Silenced and Wild-Type Plants.

(A) and (B) Mean 6 SE JA-Ile concentrations (A) and TPI levels (B) in

leaves of four to five replicate EV and TDVIGS plants, each inoculated

with Agrobacterium harboring TRV constructs, which contain either an

EV or a 335-bp TD fragment (TDVIGS). Fourteen days after inoculation,

leaves were wounded, treated with 20 mL of either M. sexta OS or OS

containing 0.625 mmol of Thr or Ile, and then harvested 1 h after

treatment for JA-Ile measurement. These plants were supplemented

daily by spraying leaves with either water or 0.25 M Thr or Ile. Three days

after OS treatment, newly hatched M. sexta larvae were placed on these

plants. TPI activity was measured after 12 d of herbivore feeding.

Asterisks represent significant differences between members of a pair

(unpaired t test: ** P < 0.01; **** P < 0.0001). FW, fresh weight.

(C) Mean 6 SE mass of M. sexta larvae after 12 d of feeding on 16 to 19

replicate EV and TDVIGS plants treated as described for (A). Asterisks

represent significant differences between members of a pair (unpaired t

test: * P < 0.05; ** P < 0.01).

(D) Mean 6 SE a-KB levels in M. sexta frass from larvae that fed on EV

and TDVIGS plants. Frass was collected from third- and fourth-instar

larvae feeding on plants treated as described for (A). Asterisks represent

significant differences between members of a pair (unpaired t test:

* P < 0.05).

3314 The Plant Cell

JA-Ile in defense signaling. The VIGS experiments in N. attenu-

ata, in which TD activity could be strongly silenced in a devel-

opmentally normal plant, reconfirmed TD’s involvement in Ile

synthesis and indicated that JA-Ile conjugation is limited by the

supply of Ile in wounded tissues (Figure 4) and that JA-Ile

regulated direct defenses and herbivore resistance (Figure 5).

It has long been known that JA is metabolized to its volatile

counterpart, MeJA, and numerous conjugates with O-glucosides,

hydroxylation, and amino acids (Sembdner and Parthier, 1993;

Sembdner et al., 1994). The glycosylated forms and amino acid

derivatives have been viewed as mere conjugates of JA, which

may be important for hormone homeostasis. Because all of the

applied conjugates could be deesterified to JA, JA and JA

conjugates were thought to have the same effect (Schaller et al.,

2004). However, recent reports have demonstrated that JA con-

jugates have their own activities. Transgenic Arabidopsis plants

that constitutively express an S-adenosyl-L-Met:JA carboxyl

methyltransferase expressed JA-responsive genes, including

VSP and PDF1.2. Furthermore, the transgenic plants showed

enhanced resistance to the virulent fungus Botrytis cinerea (Seo

et al., 2001). Studies that have applied synthetic JA–amino acid

conjugates to plants suggest that the spheres of activity within

JA–amino acid conjugates differ widely. For example, treatment

of barley (Hordeum vulgare) leaves with JA-Ile elicits JA-induced

protein without Ile cleavage from JA (Kramell et al., 1997). JA-Ile,

JA-Phe, and JA-Leu conjugates elicit accumulation of the flavo-

noid phytoalexin, sakuranetin, in rice (Oryza sativa) leaves, but

JA-Trp does not (Tamogami et al., 1997). However, it had not

been previously appreciated that JA conjugates had elicitor-

induced dynamics that were comparable to those of JA and that

subtle changes in these dynamics were associated with changes

in defense function.

The pioneering work of Staswick and colleagues (Staswick

et al., 2002; Staswick and Tiryaki, 2004) has demonstrated that

the JA-responsive gene in Arabidopsis (JAR1) adenylates JA’s

carboxyl group and that adenylated JA is actively conjugated

with various amino acids, of which Ile is quantitatively the

most important. The mutant defective in JAR1 (jar1-1) exhibits

decreased resistance to the soil fungus Pythium irregulare

(Staswick et al., 1998), implying that JA-Ile is involved in path-

ogen resistance. The analysis of JAR4VIGS plants demonstrated

that JAR4 is involved in JA-Ile conjugation (Figure 7) and that

JAR4VIGS plants are susceptible to attack by M. sexta (Figure 8),

indicating that JA-Ile is involved in herbivore resistance in

N. attenuata. Further analyses of other JA–amino acid conju-

gates at the attack site will be required to determine whether

other JA conjugates are equally as dynamically elicited and

whether these conjugates also elicit specific developmental and

defense responses in the plant.

Reduced levels of JA-Ile in asTD and TDVIGS plants resulted in

reduced levels of TPI and nicotine. Treatment of plants with JA

and JA-Ile elicited different TPI and nicotine responses, which

may be attributable to different absorption and transport rates or

to the different elicitation activities of these chemicals. The JA-Ile

burst can account for ;13% of the elicited JA burst (Figures 3Band 3C). That the quantities of JA-Ile are smaller than those of JA

may be attributable more to a rapid metabolism of JA-Ile to

unknown structures than to the conversion of JA to JA-Ile.

Alternatively, JA may be converted to MeJA or other conjugates.

The rapid declines in JA and JA-Ile may be attributable to their

binding to putative receptor(s), which have not yet been identified.

Identification of the JA receptor(s), when it occurs, will be a

breakthrough that will clarify the structural basis for the differences

in levels as well as the contents of these dynamic metabolites.

The role of JA in systemic signaling was recently demonstrated

in an elegant set of reciprocal grafting experiments. Li and

coworkers (2002, 2003) grafted the JA biosynthetic mutant spr-2,

known to be defective in fatty acid desaturase required for

JA biosynthesis, onto the JA response mutant jai-1, known to

be defective in a homolog of the Arabidopsis CORONATINE-

INSENSITIVE1 gene (Xie et al., 1998), in different combinations

and analyzed the resulting wound-induced expression of the

proteinase inhibitor II gene. Their results demonstrated that the

JA biosynthetic pathway was required to produce the long-

distance signal, suggesting that JA or related compounds de-

rived from the octadecanoid pathway function as systemically

transmitted signals in tomato. In various Nicotiana species,

nicotine synthesis in the roots is activated by leaf wounding. In

N. attenuata, this systemic response is known to require JA

signaling (Halitschke and Baldwin, 2003). JA-Ile elicits a larger

accumulation of nicotine compared with JA (Figure 5A). This

finding suggests that JA-Ile is the long-distance signal that elicits

nicotine in roots. However, JA-Ile was not detected in roots when

leaves were treated with OS, JA, or JA-Ile (data not shown),

indicating that subsequent metabolites of JA-Ile or unknown

molecules elicited by JA-Ile may be involved.

When attacked by herbivores, N. attenuata plants produce JA

and activate TD in the attacked tissues. Ile synthesized from Thr

by TD is conjugated with JA by JAR4. The resulting JA-Ile elicits

the accumulation of the direct defenses, TPI and nicotine, which

Figure 10. Proposed Roles of TD in Herbivore Resistance.

Two roles are proposed: (1) as an antinutritive defense that decreases

Thr availability in the digestive tracts of herbivores after ingestion; (2) as a

mediator of JA signaling, by supplying Ile for conjugation with JA at the

wound site and subsequently eliciting various direct defenses. JA

biosynthetic enzymes and a JA–amino synthetase are boxed. Dashed

arrows represent signal transduction pathways. LOX, lipoxygenase;

AOS, allene oxide synthase; AOC, allene oxide cyclase; OPR, 12-oxo-

phytodienoic acid reductase; JAR1, JA–amino synthetase.


in turn decrease the performance of Nicotiana-adapted herbi-

vores. The fact that the herbivore resistance of both TD- and

JAR4-silenced plants could be restored by treatment with signal

quantities (0.625 mmol/plant) of JA-Ile suggests that TD’s defen-

sive role can be largely attributed to its role in defense signaling.

When leaves of TD-silenced plants were supplemented with large

nutritional quantities (2.1 mmol/plant) of Thr, larval performance

increased. This result, in conjunction with the demonstration that

TD levels in frass of larvae that fed on water-treated TD-silenced

plants were 54% of those in frass of larvae that fed on water-

treated EV plants, suggests that TD could play a defensive role by

reducing the availability of Thr to feeding larvae. However, it

remains unclear how often the dietary Thr levels are sufficiently

high for plants to realize a defensive benefit of delivering TD

to the midgut of larvae, because resistance of the JAR4- and

TD-silenced plants was similarly impaired and both could be

restored by JA-Ile treatment (Figure 8A).

Supplementation of Thr or Ile in plants with different levels of

TD activity helped us to evaluate TD’s antinutritive role. Providing

additional Thr to leaves had positive effects on herbivore per-

formance in TD-silenced plants. However, the insufficient supply

of Ile in these plants impaired JA-Ile–mediated signaling, and

providing these plants with supplemental Ile restored their direct

defenses and TPI levels and increased herbivore resistance. Our

data also demonstrate that in normal TD-expressing plants, Ile

supplementation benefited larvae. One possible explanation is

that extra Ile provides feedback that inhibits TD activity in plants

or in the insect midgut. We plan to explore the molecular mech-

anisms involved in the regulation of TD activity in the plant as well

as in the insect midgut, following the lead of Chen et al. (2005),

who demonstrated that the regulatory domain of tomato TD

is missing in the TD protein isolated from insect midgut and

frass; this domain is responsible for feedback inhibition by Ile in

tomato.

To summarize, we propose dual roles for TD in herbivore

defense: JA-Ile–mediated signaling and antinutritive defense

(Figure 10). When attacked by herbivores, plants produce JA

and activate TD in the attacked tissues. Ile synthesized from Thr

by TD is conjugated with JA by JAR4. The resulting JA-Ile elicits

the accumulation of the direct defenses, TPI and nicotine. TD

also plays additional defensive roles by limiting the supply of

amino acids for herbivore growth when leaves are ingested by

herbivores.

METHODS

Materials and Growth Conditions

An inbred genotype of Nicotiana attenuata (synonymous with N. torreyana;

Solanaceae), originally collected from southwestern Utah in 1988, was

transformed and used for all experiments. Seeds were sterilized and

germinated as described previously (Krügel et al., 2002). Ten-day-old

seedlings were planted into soil in Teku pots and, once established,

transferred to 1-liter pots in soil and grown in the glasshouse at 26 to 288C,

under 16 h of light supplemented by Philips Sun-T Agro 400 Na lights.

Frass was collected from third- and fourth-instar Manduca sexta larvae

that fed actively during the day of the harvest and then were frozen in

liquid nitrogen and stored at �708C.

Chemical Synthesis and Treatments

JA-Ile was synthesized as described previously (Kramell et al., 1988). The

leaf undergoing the source–sink transition was designated as growing at

node 0. M. sexta larval OS were collected with Teflon tubing connected to

a vacuum and stored under argon at �808C. For OS-treated plants, theleaf growing at node þ1, which is older by one leaf position than thesource–sink transition leaf, was wounded by rolling a fabric pattern wheel

over the leaf surface to produce standardized puncture wounds. Imme-

diately after wounding, the wounds were treated with 20 mL of water, OS

at a 1:5 dilution with water, or OS containing 0.625 mmol of L-Thr, L-Ile,13C4-labeled Thr, or 13C6-labeled Ile. Leaves from JA- or JA-Ile–treated

plants were wounded with a fabric wheel and directly treated with

0.625 mmol of JA or JA-Ile. Leaves from MeJA-treated plants were

treated with 150 mg (0.625 mmol) of MeJA in 20 mL of lanolin paste as

described previously (Halitschke et al., 2000). For continuous amino acid

supplementation treatments, 0.25 M Thr or Ile in water was sprayed daily

onto the leaves on which larvae were feeding.

Generation and Characterization of asTD Transgenic Lines

For the plant transformation vector, a 1349-bp portion of the N. attenuata

TD cDNA resident on plasmid pTD13 (Hermsmeier et al., 2001) was

amplified by PCR using primers 59-GCGGCGCCATGGCATAGGTCCCA-

CAAGTTCGC-39 and 59-GCGGCGGGTCACCTGGAAGTTCTTTGTCAA-

GCC-39. The obtained 1.4-kb PCR fragment was cut with BstEII and

partially cut with NcoI. The resulting 1.4-kb fragment was cloned in

pNATGUS3 (Krügel et al., 2002) and digested with the same enzymes,

resulting in plant transformation vector pNATTD1 (10.1 kb), which con-

tained in its T-DNA a 1.4-kb fragment of TD in the antisense orientation

under the control of the 35S promoter of the Cauliflower mosaic virus. The

Agrobacterium tumefaciens (strain LBA 4404)–mediated transformation

procedure and the transformation vector have been described (Krügel

et al., 2002). Progeny of homozygous plants were selected by nourseo-

thricin resistance screening and screened for the desired phenotype,

namely, reduced MeJA-induced a-KB accumulation. For all experiments,

T2 homozygous lines, each harboring a single insertion, which was

confirmed by DNA gel blot analysis (see Supplemental Figure 2 online), or

wild-type plants were used.

JAR4 Full-Length cDNA Isolation

A cDNA fragment was obtained by RT-PCR from total RNA isolated from

wild-type plants 60 min after source leaves had been wounded with a

fabric pattern wheel. The primers were designed from the conserved

regions of Arabidopsis thaliana JAR1 and tomato (Solanum lycopersicum)

BT013679 cDNA sequences. The forward primer was 59-TTCACCTA-

TTCTTACTGG-39, and the reverse primer was 59-ACATTACTAGACAG-

TATTTGGA-39. Full-length cDNA was isolated using the GeneRacer kit

(Invitrogen) according to the manufacturer’s instructions. The 59 primer

and 59 nested primer were 59-AGAACACCTTCCCTTATATTGGTCA-

CAA-39 and 59-ACTTAAGGAAATAGTGGTAATAGGCTTT-39, respec-

tively. The 39 primer was 59-AAAGTGAATGCAATTGGAGCACTTGA-39.

Generation and Characterization of VIGS Plants

PCR was used to generate TD and JAR4 fragments from N. attenuata in

the antisense orientation with the following primer pairs: TD forward

primer, 59-GCGGCGGGATCCGCACCAAATGGCTCAACTCC-39; TD re-

verse primer, 59-GCGGCGGTCGACGTCATGCCTGTTACCACACC-39;

JAR4 forward primer, 59-GCGGCGGTCGACGTAATATTTGGCCCTGA-

TTTCC-39; JAR4 reverse primer, 59-GCGGCGGGATCCAATTGCTTAAC-

CGGCTG-39. The obtained TD (335 bp) and JAR4 (292 bp) PCR frag-

ments were digested with BamHI and SalI. The resulting fragments were

3316 The Plant Cell

cloned into the pTV00 vector digested with the same enzymes. The

pTV00 vector is a 5.5-kb plasmid with an origin of replication for Esch-

erichia coli and A. tumefaciens and a gene for kanamycin resistance

(Ratcliff et al., 2001). The A. tumefaciens (strain GV3101)–mediated trans-

formation procedure was described previously (Saedler and Baldwin,

2004). To monitor the progress of VIGS, we silenced phytoene desatur-

ase, a gene that oxidizes and cyclizes phytoene to a- and b-carotene.

These are subsequently converted into the xanthophylls of the antenna

pigments of the photosystems of plants, resulting in the visible bleaching

of green tissues (Saedler and Baldwin, 2004). When the leaves of

phytoene desaturase–silenced plants began to bleach (6 weeks after

germination; see Supplemental Figure 9 online), leaves of TD-silenced

(TDVIGS), JAR4-silenced (JAR4VIGS), and empty vector–inoculated (EV)

plants were used.

Nucleic Acid Blot Analysis

Extraction of total RNA and RNA gel blot analysis were performed as

described previously (Winz and Baldwin, 2001). Genomic DNA was ex-

tracted from leaves as described previously (Richard, 1997), and 10 mg of

DNA was digested with EcoRI and blotted onto nylon membranes.

To prepare the probe, plasmid pTD13 (GenBank accession number

AF229927) containing the full-length cDNA of TD was cut with PstI and

gel-eluted using the Geneclean kit (BIO 101), labeled with 32P using the

RediPrime II random prime labeling kit (Amersham-Pharmacia), and

purified on G50 columns (Amersham-Pharmacia). After overnight hybrid-

ization, blots were washed three times with 23 SSPE (13 SSPE is 0.115

M NaCl, 10 mM sodium phosphate, and 1 mM EDTA, pH 7.4) at 428C and

one time with 2 3SSPE and 2% SDS at 428C for 30 min and then analyzed

on a phosphor imager (model FLA-3000; Fuji Photo Film Co.).

Real-Time PCR Assay

Total RNA was extracted with TRI reagent (Sigma-Aldrich) according to the

manufacturer’s instructions, and cDNA was prepared from 200 ng of total

RNA with MultiScribe reverse transcriptase (Applied Biosystems). The

primers and probes specific for TD and JAR4 mRNA expression detection

by quantitative PCR were as follows: TD forward primer, 59-TAAGG-

CATTTGATGGGAGGC-39; TD reverse primer, 59-TCTCCCTGTTCACGA-

TAATGGAA-39; JAR4 forward primer, 59-ATGCCAGTCGGTCTAACT-

GAA-39; JAR4 reverse primer, 59-TGCCATTGTGGAATCCTTTTAT-39; ECI

forward primer, 59-AGAAACTGCAGGGTACTGTTGG-39; ECI reverse primer,

59-CAAGGAGGTATAACTGGTGCCC-39; FAM-labeled TD probe, 59-TTT-

TTAGATGCTTTCAGCCCTCGTTGGAA-39; FAM-labeled JAR4 probe,

59-CAGGTCTGTATCGCTATAGGCTCGGTGATGT-39; FAM-labeled ECI

probe, 59-CGTCAAAATTCTCCACTTGTTTCAACTGT-39. The assays us-

ing a double dye-labeled probe were performed on an ABI Prism 7700

sequence detection system (qPCR Core kit; Eurogentec) with N. attenu-

ata sulfite reductase (ECI) for normalization and according to the man-

ufacturer’s instructions with the following cycle conditions: 10 min at

958C; then 40 cycles of 30 s at 958C and 30 s at 608C.

TD Activity Measurement

Leaves or M. sexta frass were homogenized in 2 volumes of extraction

buffer (100 mM Tris buffer, pH 9, 100 mM KCl, and 10 mM b-mercap-

toethanol) and centrifuged at 15,000g for 15 min at 48C. TD activity was

assayed by incubating the enzyme with substrate and determining the

quantity of a-KB formed. The a-KB was estimated by modifying the

method described by Sharma and Mazumder (1970). Protein extract

(100 mL) was added to the same volume of reaction buffer (40 mM L-Thr,

100 mM Tris buffer, pH 9, and 100 mM KCl). After incubation at 378C for

30 min, 160 mL of 7.5% trichloracetic acid was added to stop the reaction,

and the protein precipitate was removed by centrifugation at 10,000g for

2 min. The a-KB was determined by adding 400 mL of 0.05% dinitrophe-

nylhydrazine in 1 N HCL to the sample solution. After incubation at room

temperature for 10 min, 400 mL of 4 N sodium hydroxide was added to the

sample solution and mixed well. After incubation at room temperature for

20 min, the absorbance of the sample solution was read at 505 nm in a

spectrophotometer (model Ultraspec 3000; Pharmacia Biotech).

M. sexta Performance

Leaves at nodes þ1 and þ2 were wounded and treated with JA or JA-Ileor left untreated. For the effects of TD on M. sexta larval mass in untreated

and JA- and JA-Ile–treated transgenic and wild-type plants, freshly

hatched larvae (North Carolina State University) were placed on 7 to 16

replicate leaves at node 0 on individual plants, 3 d after treatment. Larval

mass was measured at 2, 4, and 6 d or at 3, 6, and 9 d after larvae were

allowed to feed on the plants. In the experiments with VIGS plants, freshly

hatched larvae were placed on 12 to 19 replicate leaves (on separate

plants), 3 d after elicitation. Larval mass was measured at 6, 9, and 12 d

after larvae began feeding.

Analysis of Direct Defense Traits

Nicotine, chlorogenic acid, and diterpene glycoside were analyzed by

HPLC as described previously (Keinanen et al., 2001) with the following

modification of the extraction procedure: ;100 mg of frozen tissue washomogenized in 1 mL of extraction buffer using the FastPrep extraction

system (Savant Instruments). Samples were homogenized in FastPrep

tubes containing 900 mg of lysing matrix (BIO 101) by shaking at 6.0 m/s

for 45 s.

TPI activity was analyzed by radial diffusion activity assay as described

previously (van Dam et al., 2001).

JA and JA-Ile Measurement

Leaves were harvested and immediately frozen in liquid nitrogen. Sam-

ples were homogenized in 3 volumes of extraction buffer (acetone:50 mM

citric acid, 7:3 [v/v]). Samples were centrifuged at 13,000 rpm for 15 min at

48C, and supernatants were transferred to a new tube. The pellet was

reextracted with extraction buffer. The combined supernatants were

evaporated to dryness in a heating block, and the remaining aqueous

phase was extracted three times with 1 mL of ether. The ether layer was

evaporated completely and the residue dissolved in acetonitrile. The

samples were separated by an Agilent LC1100 HPLC system with de-

gasser, binary pump, autoinjector, and column thermostat and detected

by a diode array detector coupled to a LCQ DECA XP mass spectrometer

(Thermo-Finnigan). Mobile phase A consisted of 0.5% acetic acid in water

and mobile phase B consisted of 0.5% acetic acid in acetonitrile. The

mobile phase gradient was increased linearly from 20% B (initial value) to

50% B at 16 min, held constant at 50% B for 25 min, and subsequently

increased linearly to 100% B at 30 min. The mobile phase flow was

0.7 mL/min, and the injection volume was 30 mL. The stationary phase was

a Luna 5m C18 column (250 3 4.60 mm, 5-mm particle size; Phenomenex).

The mass spectrometry conditions were as follows: atmospheric pres-

sure chemical ionization source, 5008C vaporizer temperature; 2758C

capillary temperature; 10-mA discharge current; sheath gas, nitrogen,

50 arbitrary units; auxiliary gas, nitrogen, 30 arbitrary units. Three tandem

mass spectrometry ion-acquisition segments were programmed as fol-

lows: (1) 10 to 17.5 min, m/z 155 at 28 negative polarity for 2-chloroben-

zoic acid (internal standard); (2) 17.5 to 21.5 min, m/z 211 at 23 positive

polarity for JA. The third segment (21.5 to 30 min) contained the following

three scan events: (1) m/z 324 at 30 positive polarity for endogenous

JA-Ile; (2) m/z 328 at 30 positive polarity for synthetic JA-Ile derived from

[13C4]L-Thr (Cambridge Isotope Laboratories); (3) m/z 330 at 30 positive

polarity for synthetic JA-Ile derived from [13C6]L-Ile (Cambridge Isotope


Laboratories). Standard curves were constructed with known quantities

of Ile, JA, and JA-Ile and used to quantify those chemicals in samples. The

tandem mass spectrometry spectra of JA and JA-Ile are given in Sup-

plemental Figure 10 online.

To estimate the JA and JA-Ile responses, we integrated the amount

produced in each leaf from 0 to 5 h.

Accession Numbers

Sequence data for the full-length cDNAs for N. attenuata TD and JAR4

can be found in the GenBank/EMBL data libraries under accession

numbers AF229927 and DQ359729, respectively. Accession numbers for

the sequences in phylogeneic analysis are given in the Supplemental

Methods online.

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Methods and References.

Supplemental Figure 1. Expression of TD in N. attenuata Plants

Attacked by Insect Herbivores and Elicitors.

Supplemental Figure 2. DNA Gel Blot of Genomic DNA in Wild-Type

and asTD Plants.

Supplemental Figure 3. Comparison of Growth Rates of Wild-Type

and T2 asTDM Plants Grown in Individual Pots or Competing with

Each Other in the Same Pot.

Supplemental Figure 4. Silencing TD in asTD and TDVIGS Plants

Improves Herbivore Performance.

Supplemental Figure 5. Chlorogenic Acid and Diterpene Glycoside

Concentrations Elicited by JA and JA-Ile Treatments to Leaves in

Wild-Type and asTDM2 Plants.

Supplemental Figure 6. Phylogenic Tree of the JAR Family Proteins.

Supplemental Figure 7. Alignment of Deduced Amino Acid Se-

quences of JARs from Nicotiana attenuata (JAR4), Arabidopsis

thaliana (JAR1), Solanum lycopersicum (BT013697), Oryza sativa

(GH3.5), and Nicotiana glutinosa (BAE46566).

Supplemental Figure 8. DNA Gel Blot of JAR4 in Wild-Type Plants.

Supplemental Figure 9. VIGS Plants during the Stalk Elongation

Stage.

Supplemental Figure 10. Analysis of JA and JA-Ile Conjugates by

LC-MS.

ACKNOWLEDGMENTS

We thank Bernd Krock for unflagging assistance in LC-MS analysis and

for the synthesis of the JA-Ile; Thomas Hahn for sequencing; Klaus

Gase, Susan Kutschbach, and Wibke Kröber for the construction of

vectors pNATTD1, TDVIGS, and JAR4VIGS; Tamara Krügel and Michell

Lim for the plant transformation; Anke Steppuhn for assistance in the

HPLC analysis; Dominik Schmidt for assistance in the real-time PCR

analysis; Rayko Halitschke for help with data analysis and for the first

sequences of JAR4; and Emily Wheeler for editorial assistance. This

work was supported by the Max-Planck-Gesellschaft. A.G. acknowl-

edges the Alexander von Humbolt Foundation (Bonn, Germany) for a

research fellowship.

Received January 13, 2006; revised August 25, 2006; accepted October

13, 2006; published November 3, 2006.

REFERENCES

Azevedo, R.A., Arruda, P., Turner, W.L., and Lea, P.J. (1997). The

biosynthesis and metabolism of the aspartate derived amino acids in

higher plants. Phytochemistry 46, 395–419.

Baldwin, I.T. (1998). Jasmonate-induced responses are costly but

benefit plants under attack in native populations. Proc. Natl. Acad.

Sci. USA 95, 8113–8118.

Chen, H., Wilkerson, C.G., Kuchar, J.A., Phinney, B.S., and Howe,

G.A. (2005). Jasmonate-inducible plant enzymes degrade essential

amino acids in the herbivore midgut. Proc. Natl. Acad. Sci. USA 102,

19237–19242.

Colau, D., Negrutiu, I., Van Montagu, M., and Hernalsteens, J.P.

(1987). Complementation of a threonine dehydratase-deficient Nico-

tiana plumbaginifolia mutant after Agrobacterium tumefaciens-

mediated transfer of the Saccharomyces cerevisiae Ilv1 gene. Mol.

Cell. Biol. 7, 2552–2557.

Dammann, C., Rojo, E., and Sanchez-Serrano, J.J. (1997). Abscisic

acid and jasmonic acid activate wound-inducible genes in potato

through separate, organ-specific signal transduction pathways.

Plant J. 11, 773–782.

Glawe, G.A., Zavala, J.A., Kessler, A., Van Dam, N.M., and Baldwin,

I.T. (2003). Ecological costs and benefits correlated with trypsin

protease inhibitor production in Nicotiana attenuata. Ecology 84,

79–90.

Halitschke, R., and Baldwin, I.T. (2003). Antisense LOX expression

increases herbivore performance by decreasing defense responses

and inhibiting growth-related transcriptional reorganization in Nicoti-

ana attenuata. Plant J. 36, 794–807.

Halitschke, R., Gase, K., Hui, D.Q., Schmidt, D.D., and Baldwin, I.T.

(2003). Molecular interactions between the specialist herbivore Ma

Silencing Threonine Deaminase and in Nicotiana attenuata ... › content › plantcell › 18 › 11...

Documents

Transcript of Silencing Threonine Deaminase and in Nicotiana attenuata ... › content › plantcell › 18 › 11...